The RTA Walkthrough and Use Thread. With video. By Erin H.

[Salami]

I am copying this from MSS with Erin's permission. http://www.mobilesoundscience.com/sh...ead-With-video

This is intended to lay out the basics of measuring your car audio system using Room EQ Wizard, which can be found here:
REW - Room EQ Wizard Home Page


Here goes...




As mentioned above, I use REW for my testing. It's free and pretty easy to use once you figure out how to use it. :blush:
I've used every bit of software under the sun. At this point, I've just grown to like REW more than I used to so I've gone back to it for my tuning measurements. For those wanting to see how I'm using REW, here's some info.

RTA vs Impulse Response Measurement Method:
One thing to note is there are different meanings of the term, at least how we use it:

1. RTA - Real Time Analyzer:
   
   • This is simply a real time measurement of what the mic hears. Birds chirping, subs playing... whatever. It records it.
   • RTAs are typically used to record pink noise.
   
   
2. Impulse Response:*
   
   • This can be a form of RTA, depending on how you look at it. An impulse is used typically to measure something before a reflection because you can gate the response. In other words, let's say I want to measure Speaker A. I know that the walls and floors create reflections occurring after 3 milliseconds (ms) that will 'tarnish' the speaker's response as measured by the mic if I let it. To keep this from happening, I look at the impulse response, tell the software to ignore everything after 3ms. Bammo... no more reflections in the measurement. Just Speaker A.
   • The impulse response is measured by sweeping a sine wave and capturing the response.
   
   

* I have severely watered down my explanations here and there are caveats; especially when you get in to different window type methods. But for the sake of this post, it's fine.

The bottom line: In a car, you don't care about impulse gating. You can't really achieve a reflection free zone so there's not much point in trying. Therefore, we just disregard the whole gating process. That leaves us with a very long impulse response that matches what an RTA would show you, if the signal were the same (ie: pink noise). IOW, using a very long impulse window (100's of milliseconds) in the car will yield an RTA measurement.

Why use this impulse method if it essentially nets you the same thing as an RTA measurement? Because RTA measurement only gives you RTA data; SPL vs frequency. You can't get Decay or some of the other things I am looking to get. More data. That's all. I get in to it more below.



Cliffs:

• When measuring a car, there are a couple ways to do it.
• RTA and Impulse are not the same. They each have their own use. However, when the impulse is used without filtering or gating it, it nets you the same result as an RTA.
• The benefit of using impulse measurements are you get more data such as decay, group delay, etc.
• Like a standard RTA measurement, multiple impulse responses should be taken and averaged together if you want to tune to a car.

[Salami]

I'll discuss how to use either of the two methods listed above. But first off, let's discuss the equipment you need and why you need it.

• If doing RTA only measurements all you really need is:
  
  1. A CD with pink noise
  2. A mic to capture the sound.
  
  
• If doing impulse measurements you need two things:
  
  1. A way to use the software's signal generator and send that signal to your audio system. This can be done by using a soundcard output and run it in to your audio system auxiliary input.
  2. A mic is used to capture the sound.

If you are looking to achieve a target curve of any sort you MUST USE A CALIBRATION FILE with the mic.



-------------------------------------


For impulse measurements, this is the gear I'm using. It's pretty simple:

1. M-Audio Transit. This sends the signal from REW software via a 3.5mm male/female cable to my P99's auxiliary input. I use this because I don't have on-board audio.
2. Dayton Omnimic USB microphone. It comes with a cal file.

You don't have to use what I have. In fact, you can save a good deal of money by simply using:

1. Your laptop's built-in headphone output
2. This Dayton mic (click link here). There are numerous other mic alternatives. This is just an easy one with one USB cable. Plug and chug. And it comes with cal files.

------------------------------------------


You know what you have to have. Why do you have to have it?

1. The soundcard output is used to send the signal to your auxiliary source. Every headunit should now have one of these. This is how I do my testing most of the time. If it doesn't, you may be limited to using the pure RTA method only.
2. The mic records the system output. Simple as that.

-----------------------------------------


Here's some pictures of my gear:


M-Audio Transit, using the output via a 3.5mm headphone extension cable.








Dayton Omnimic USB mic (it has electrical tape because I broke the tip; these things are pretty fragile and I made a goof):







Mic in the headrest:









3.5mm aux cable plugged in to my headunit's aux input on the face panel:

...

[Salami]

Woke up early before the little one did to make this video so forgive me if I sound half dead and my video has the shakes.

This covers only the RTA aspect of using REW. I'll post another video soon showing how to use the impulse measurement method since I feel it's a bit more complete depending on what you want to do and given I'll be posting results from that kind of measurement.

Let me know if you have questions.



[youtube]bds759UuWiU[/youtube]



- Erin

...

[Salami]

Erin pertaining to pink noise, what kind should be used?

From what I can see there are at least 3 kinds. If the system as a whole is being RTA'd, should uncorrelated pink noise be used or does it not matter?


Thanks.

[littlemissGTO]

Thanks for posting this. After the GTG I attended it makes so much more sense.

[highly]

...enter the discussion about spectral averaging to achieve a uniform soundfield measurement for capturing data above 200Hz in a confined space using RTA software...

[erinh]

Erin pertaining to pink noise, what kind should be used?

From what I can see there are at least 3 kinds. If the system as a whole is being RTA'd, should uncorrelated pink noise be used or does it not matter?


Thanks.

For RTA? I'd just go with correlated (mono).

You can pan it through your headunit's fader or simply mute the channels you don't want from your DSP if you want to measure a single side at a time.